Motion compensated inter-frame coding has been widely adopted in various coding standards, such as MPEG-1/2/4 and H.261/H.263/H.264/AVC. While motion-compensated inter-frame coding can effectively reduce bitrate for compressed video, Intra coding is required to compress the regions with high motion or scene changes. Besides, Intra coding is also used to process an initial picture or to periodically insert I-pictures or I-blocks for random access or for alleviation of error propagation. Intra prediction exploits the spatial correlation within a picture or within a picture region. In practice, a picture or a picture region is divided into blocks and the Intra prediction is performed on a block basis. Intra prediction for a current block can rely on pixels in neighboring blocks that have been processed. For example, if blocks in a picture or picture region are processed row by row first from left to right and then from top to bottom, neighboring blocks on the top and neighboring blocks on the left of the current block can be used to form Intra prediction for pixels in the current block. While any pixels in the processed neighboring blocks can be used for Intra predictor of pixels in the current block, very often only pixels of the neighboring blocks that are adjacent to the current block boundaries on the top and on the left are used.
The Intra predictor is usually designed to exploit spatial features in the picture such as smooth area (DC mode), vertical line or edge, horizontal line or edge and diagonal line or edge. Furthermore, spatial correlation often exists between the luminance (luma) and chrominance (chroma) components. Therefore, reconstructed luma pixels can be used to derive the Intra chroma prediction. In the emerging High Efficiency Video Coding (HEVC), a chroma Intra prediction mode based on the reconstructed luminance signal has been considered. This type of chroma Intra prediction is termed as Linear Model (LM) prediction. FIG. 1 illustrates the Intra prediction derivation for LM mode. First, the neighboring reconstructed pixels (indicated by circles) of a collocated luma block (i.e., Y block) and the neighboring reconstructed pixels (indicated by circles) of a chroma block (i.e., U or V block) in FIG. 1 are used to derive the linear model parameters between the blocks. The predicted pixels of the chroma block are generated using the parameters and the reconstructed pixels of the luma block. In the parameters derivation, the top reconstructed pixel row adjacent to the top block boundary of the current luma block and the left reconstructed pixel column adjacent to the left block boundary of the current luma block are used. It is noted that the second left reconstructed pixel column from the left boundary is used instead of the left column immediately adjacent to the left boundary in order to match the sampling locations of the chroma pixels. The specific row and column of the luma block are used in order to match the 4:2:0 sampling format of the chroma components. While FIG. 1 illustrates the example of LM chroma mode for the 4:2:0 sampling format, the LM chroma mode for other chroma sampling format may also derived similarly.
According to the LM prediction mode, the chroma values are predicted from reconstructed luma values of a collocated block. The chroma components may have lower spatial resolution than the luma component. In order to use the luma signal for chroma Intra prediction, the resolution of the luma signal may have to be reduced to match with that of the chroma components. For example, for the 4:2:0 sampling format, the U and V components only have half of the number of samples in vertical and horizontal directions as the luma component. Therefore, 2:1 resolution reduction in vertical and horizontal directions has to be applied to the reconstructed luma samples. The resolution reduction can be achieved by down-sampling process or sub-sampling process.
In LM chroma mode, for a to-be-predicted chroma sample V with its collocated reconstructed luma sample Vcol, the linear model to generate LM predictor P is formulated as follows:P=a·Vcol+b 
In the above equation, a and b are referred as LM parameters. The LM parameters can be derived from the neighboring reconstructed luma and chroma samples around the current block so that the parameters do not need to be coded in the bitstream. After deriving the LM parameters, chroma predictors can be generated from the collocated reconstructed luma samples in the current block according to the linear model. For example, if the video format is YUV420, then there are one 8×8 luma block and two 4×4 chroma blocks for each 8×8 coding unit, as shown in FIG. 1., In FIG. 1, each small square corresponds to one pixel in the current coding unit (2N×2N for luma and N×N for chroma) to be coded. The LM parameters are derived first based on neighboring reconstructed samples of the current coding unit, which are represented as circles in FIG. 1. Due to the YUV420 sampling format, the collocated chroma position is located between two corresponding vertical luma samples. An average value between two corresponding vertical luma samples is used to derive the LM parameters. For neighboring pixels above the top block boundary, the average value is replaced by the closest sample in the vertical direction in order to reduce the line buffer requirement. The neighboring pixels (as shown in circles) of the currently luma (Y) and chroma (U or V) coding units are used to derive the LM parameters for the respective chroma component as shown in FIG. 1. After the LM parameters are derived, the chroma predictors are generated based on the linear model and the collocated luma reconstructed samples. According to the video format, an average luma value may be used instead of the corresponding luma sample.
A method of chroma Intra prediction using extended neighboring pixels for LM parameter derivation has been disclosed by Zhang et al., (“New Modes for Chroma Intra Prediction”, in Joint Collaborative Team on Video Coding (JCT-VC) of ITU-T SG16 WP3 and ISO/IEC JTC1/SC29/WG11, 7th Meeting: Geneva, CH, 21-30 Nov., 2011, document: JCTVC-G358). FIG. 2A-FIG. 2C illustrate an example of chroma Intra prediction for 8×8 chroma block using extended neighboring pixels according to Zhang. FIG. 2A corresponds to regular chroma Intra prediction being considered by HEVC. FIG. 2B illustrates the example of LM parameter derivation based for an additional chroma Intra mode using extended horizontal neighboring pixels, where additional N pixels from the upper-right neighbor are used. FIG. 2C illustrates the example of LM parameter derivation based for another additional chroma Intra mode using extended vertical neighboring pixels, where additional N pixels from the lower-left neighbor are used. While the method of Zhang demonstrates noticeable improvement in performance, the method also causes increases in computational complexity and buffer requirement.
It is desirable to develop improved method that may further improve the performance and/or reduce the buffer requirement of chroma Intra prediction.